How likely are junior analysts getting replaced by AI?
I work in investment banking but have always been interested in hedge funds. Curious to gather opinions or hear interesting experiences.
It seems like firms and the big dogs have different takes on AI. Ken said Gen AI can’t deliver real alpha, while Viking is deploying VikingGPT with usage reportedly up 400% YoY. I understand the latter doesn’t necessarily mean Gen AI creates alpha, but it does seem to help analysts and PMs get on the right track faster.
So my question is:
- How likely are junior analysts at hedge funds (especially pods) to be replaced by AI, given that technology can now streamline research, test trade ideas, analyze risk/reward, and even build models in the near future?
- When might this happen—and what’s the value add for junior analysts who can stay relevant and continue to be hired?
Any insights are much appreciated. Thank you.
To a large extent the purpose of an analyst is not analysis but instead to confer prestige on the organization. This is much of the purpose of your work - you symbolically perform merit and struggle. This is something that I do not think an AI can replace. Status is a positional good - it requires rejecting some folks. Ai is sort of infinite and therefore cannot grant status in the same way. This is a reason why it is not very useful for hedge funds.
That said, I do think the substantive work of an analyst can be performed by an AI in 5 years or so. The main barrier is the cost of obtaining the raw data about what analysts do but once that barrier is surpassed then it is a relatively trivial learning problem. There is enough going on that these economic costs will be surpassed.
However saying that the work of a hedge fund analyst can be performed by an AI is not the same as saying it can be replaced by an AI.
It is perfectly possible hedge funds never adopt this technology but are not replaced. Conflating the two assumes that hedge funds are super rational alpha maximizing entities which certainly they are not. In my base case scenario, there is a very narrow group of prop funds that adopt the technology and essentially take the alpha. In this scenario, hedge funds will see alpha decay but not actually change their organizational practices. The hedge funds will see performance degradation but blame other factors for that. In other words, the existing phenomena we have seen for the last 10 years will simply continue to occur.
Generally hedge funds do not change their behaviors or adopt new technology. Instead when they get old fashioned they just stop doing well but otherwise keep doing what they did before.
I’m sure funds like Viking will have cute internal tools for marketing purposes to LPs. These are really valuable initiatives but are not substantial changes to investment process. The process by which technological innovation is performed theatrically is a separate process imo. I do think this type of adoption will be very strong but this is not the same as getting rid of analysts.
agree and disagree on some points. HF Analysts will never become extinct. Firms will, however, be able to do more with less. So there will be gradual reduction in the number of seats available. The "analyst" will not be completely replaced, but there will certainly be people that are displaced and breaking into to a HF seat will be much more difficult (i.e., what used to take 2 analyst time to complete can now be done by 1). HFs, on average, are more rationale than you give them credit for. Those that fail to integrate AI will be less efficient and eventually at disadvantage. HF industry is Darwinian in that those underperform or become obsolete die off, while new firms are born. What slows this process is that capital allocations are relational at times causing semi-inefficiency. Large firms that were early to integrate AI will continue to persist. if a PM can operate at 75% analyst headcount by leveraging AI efficiently with no degradation to work output/quality, you think that the PM would maintain his current analyst headcount bc of the symbolic nature that the "analyst" represents? That's money directly out of his pocket.
HFs are still in early stages of courting AI - displacement of some of our peers will occur through by less hiring after attrition as PMs/firms continue to get more comfortable with capabilities and quality of output AI yields. I think you are discounting how rapidly it is advancing.
agreed.
I imagine a more top-heavy personnelle model once AI can model, etc.
hedge funds have been underperforming the market for the last two decades. you can look at any list of the top 100 hedge funds and find that nearly all of the best performing firms still underperform the index. i agree and disagree that it is darwinian. making a lot of money at a hedge fund is not darwinian. all you need is for an allocator to give you money. when funds underperform there is no statistic you can point to to know for sure the underperformance will persist, so there are a lot of vicissitudes that keep people in the game. moreover, hedge funds reduce correlation for their clients. so long as they are unique it's ok for them to perform poorly. the most important traits are handsome, pedigree, good at talking -- this sort of performing arts is the type of performance that matters most. (there are other valuable traits, like data manipulation, that are very valuable as well. for example, if you earn excess returns on a small capital base you can continue to point to an impressive-seeming statistic of your annualized return since inception. this trick as well as the general fantasy of the "lone genius" is enough to keep an entire industry afloat).
if you mean that alpha is darwinian you are absolutely right though. but darwinian evolution is slow. it takes a long time for a species to go extinct. ai has already been replacing hedge funds for the last two decades, but the hedge fund is resistant to this and has produced a select number of new models that are insulated against this competition (like global equities at citadel). there are plenty of funds that earn massive amounts of alpha, but they are just not hedge funds. that part of the industry is hidden because it does not advertise, raise outside money, or brag. there are operating businesses that earn high enough ROIC that they don't need to raise debt or equity, because doing so would be dilutive. similarly, there are investment firms that simply don't take outside investors becuase they are just too profitable to do so. the ONLY case where one of these entities has felt a reason to brag was RenTech. RenTech had some strategies that did not work but there was an opportunity to create a hedge fund product around selling such strategies to silly allocators. as a result, it was worth it for them to advertise the performance of their Medallion Fund. it is a privilege to simply be in a room with RenTech if you are an allocator so many were undoubtedly happy to use the hedge fund product to "reduce correlation" simply because it looks good to invest behind the same firm that created Medallion.
the general predictions about AI creating a concentration of success in the hands of the few is exactly what has already happened in public markets (e.g., perhaps 10 firms print insane returns they don't tell anyone else about and capture all the alpha there is really). in other words, the future that people fear has to some extent already happened without anyone noticing. the traditional hedge fund industry however continues to exist for fascinating anthropological reasons, despite that they have already in a sense been outcompeted in a darwinian sense. we have religious and spiritual needs to feel control over our future and, for a certain demographic, the modern hedge fund helps with this need, despite that they have to keep shifting the goal posts (from "absolute return" to now "we will say we achieve absolute return but this was a bad year though we still reduce correlation!")
i actually think that even in the most optimistic AGI / ASI scenario to some extent the story of RenTech is a story of hope for us mere mortals. even when we are outcompeted, human jobs are surprisingly resilient because humans have a remarkable ability to both create and believe mythology, to both act and suspend disbelief.
Most tiger cubs (Viking less so given they’re now run like a semi pod) strategy is finding right beta to long (they’ll deny this tho), so more so than in a pod setting, ai can replace those guys first.
“Hey get me all the companies in AI datacenters value chain”
genuinely curious on this point, what's wrong with betting right beta early on which generates absolute returns with good risk-reward ratio? (ofc not doing datacentre play in 2H25), is this not worth LP allocating to them?
Nothing wrong. But those guys are not even good at it, and terrible risk managers. If you deliver in-line during bull market, and go down more during bear market, what’s the point of LP allocating any capital to them.
Set aside what’s better vehicle for LPs, SM on average is a better job for dura/stability imo. Yes, if you have a big year at a pod, you get paid well, but making big $ in pod model isn’t as easy as riding beta.
I think this point is definitely correct, but from my view it actually makes their job harder to automate (this is coming from someone working at a pod). Their job is a judgement game more than ours and that’s something that AI is least adept at replicating.
Soft skills and human interaction will become a lot more valuable as AI improves but it will never replace analyst only change their jobs/roles
Remember, there is a difference between AI actually being competent at something and hiring managers making employment decisions based on perceived abilities of AI
I don’t know why everyone assumes that when AI can do financial analysis accurately - that the plan is to replace the most junior headcount / be more senior heavy. Senior headcount is EXPENSIVE. Wouldn’t it make sense to gut the duplicative middle management layers, hire cheap junior talent that is armed with AI so that they can effectively be the “associate” to the AI? Then train them the process side - obviously coming from a sell side advisor perspective. HFs are a different thing, idk if I think about where it might be best to cut, it’s probably the middle layer
Most hedge funds, unlike investment banking or private equity teams, are run lean.
I work at a single manager with $500+ million of fee paying AUM per analyst. Even at 1% mgmt fee, can easily support comp and overhead. Many analysts don’t even have a junior — out of choice; if they wanted one some kkr or apo associate would show up on our doorstep.
So at places like mine (and even as headcount heavy as coatue, watch the podcast) there will be no headcount reduction so long as returns are there and there aren’t net outflows. Instead many of us will be (and already are are) leveraging AI to generate more p&l, cover more sectors, peer into foreign mkts outside our core coverage, etc.
But if you’re at heavy headcount place or pe / banking, watch out. Theres no need for a VP to have an associate or analyst if AI can expedite deck and model creation.
Ex nesciunt dolor nesciunt nobis excepturi quis dolores. Aut velit ab blanditiis rem ut nemo. Quo aperiam natus sequi. Sunt similique voluptatum animi velit animi dignissimos.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...